Summary of ANSYS and Strain Gauge Results for the EC Calorimeter OH and MH Modules

Abstract

The OH and MH modules of the EC calorimeter consist essentially of metal boxes containing calorimetry plates. These plates can contribute to the module behavior only in compression, with this effect being enhanced if the plates are compressively preloaded against the skin of the box prior to assembly. The finite element method can be applied in the analysis of these modules. Its advantages are: 1. The structural components can be modeled with less simplification than beam theory allows. The angled faces of the OH modules can be represented exactly, and the shear deflections inherent in short, deep beams will be a natural part of the solution. 2. The finite element method can be subjected to any number of realistic loadings. 3. With proper mesh density relevant stresses can be extracted. The disadvantages of the method are that exact modeling of the internal plates is difficult, time consuming, and computationally expensive. It is of interest, then, to verify how well a simple model of the structural components only (i.e., the skin, endplates, and any structural internal plates) predicts deflections and stresses which can be relied on for design purposes. The finite element modeling of the OH and MH EC modules has been under constant review since the technique was first applied to these structures. Early verification attempts were based on comparison of finite element deflection predictions with measured module deflections. These comparisons were not entirely successful, due primarily, in the author's opinion, to the difficulty of measuring the actual module deflections with acceptable accuracy. It was proposed in October, 1986, that verification be based on stress, rather than deflection. The purpose of this report is to summarize the results of four experiments which were conducted to determine the accuracy with which ANSYS finite element models could predict the stresses in the OH and MH EC modules as measured by strain gauges. The three comparisons with actual module prototypes show that ANSYS can predict with good accuracy the stresses in those regions far from discontinuities where the stress gradient is low. In all regions, but particularly those of high gradient, ANSYS will tend to overestimate the stress. The comparison with the skin-only module shows that the basic approach is sound and exhibits the behavior expected from a finite element analysis. Finite element analysis can clearly be a useful part of the module design process when augmented by experimental and closed-form analytical techniques.